From: "Tom Sawyer" <transami@...>
> On Sat, 2002-09-14 at 10:56, Oren Ben-Kiki wrote:
> > Tom Sawyer wrote:
> > > i think Oren, and the lik,e have a point that using the - along with the
> > > type def, say !omap, works well enough. so do we want to go to all the
> > > trouble of implementing GCM now? for purely asthetic reasons? i think
> > > this needs a good bit of discussion still and thus should wait til
> > > version 2, so to speak, before anysort of official implementation. --but
> > > perhaps i'm wrong.
> > +1. I think this issue was blown out of proportion and stands a good chance
> > of making YAML not being YAML any more; either it turns into a host of
> > different YAMLs (Goldfish vs. whatever) or it becomes XML. Either way,
> > that's not a desirable outcome.
> > I think we've trashed this to death; the posts seem to be repeating
> > themselves. It also worries me that we don't seem to be getting any closer
> > towards closure. Also, that many points that were raised a zillion times get
> > a reponse of "cool" or "nice decision" as if they are new.
> > I don't know what's the best way to proceed...
> well, i think what happens is that threads get lost as we move from one
> thing to the next but its hard because many things interrelate.
> i think we need an offical wiki page for this that has a list of things
> in discussion/to be determine. then we can prioitize it and start
> knocking them out. do we have a pge like this?
> so to do this i think we each should spell out the issues we beleive to
> be at hand.
> i'll start my little list now:
> 1: >
> '//: goes out the window (perhaps that one is already done)
> persistent comments can be an issue for v2'
> 2: >
> we hve to decide an official notation for !omap and !dmap.
> do we use - or not? if we don't then at least some GCMish
> code will have to be used by implementors.
> 3: >
> we should officialy classify the types into three levels.
> "0th order" which has the primary three --like primary colors ;)
> then "1st order", having !set, !omap and !dmap along
> with the others, like !int, !float, etc. and then
> the 2nd class types for all the other "user" defined
> thingamawhatchies. we can determine better names for them too.
> 4: >
> determine if the 1st order types should be included
> in the spec and expected to be supported by implementors
> of offical yaml compliant processors.
> 5: >
> figure out the darn implict behavior!
> 6: >
> nail down the v1 information models.
> perhaps the most difficult task at hand.
> (i have a related question to this which i have
> been devloping in earlier posts and will
> address in a later post)
> 7: >
> after all these and and any others we can think of,
> then its time to clean up the spec and freeze v1.
> okay that's my synopsis. what's yours? add to this.
> by the way, lets not "dispute" each other synopsis.
> lets just take them them for what they are an address
> them one by one. we can clarify and add to them, but
> no trying to invalidate them UNTIL its time to address that
> issue. that make sense?
I agree with you in spirit, Tom, but let me give my spin on it.
If folks have an issue that they want to keep pushing, and they get tired of
making the same arguments, they should make a Wiki page that summarizes their
arguments. Example Wiki pages:
Wiki pages could be a particularly useful tool for Clark, who seems to send
multi-paragraph emails by the dozens. Surely there's some repetition between
those emails, which could be avoided by writing succinct Wiki pages.
If you don't feel like your proposal is getting enough of an audience from
Clark, Oren, and Brian, you might want to create a YAC page that points to your
proposal. Look at RecentChanges in the Wiki to see how the YAC system works.
See these pages in particular:
Finally, don't get discouraged if Oren, Clark, and Brian blow you off at first.
They eventually come up with a solution to most problems that I've seen. A good
example of this is their recent willingness to reorganize the spec.
If you don't let have the Wiki bookmarked, please bookmark this page:
If you look today, for example, you will see two very cool schema proposals from
Why and Tom.